← Back to glossary Browse letter H hub

Hallucinations

Confident but incorrect outputs generated by AI models.

Definition

Confident but incorrect outputs generated by AI models.

How It Works

Hallucinations helps teams build predictable AI and translation workflows by setting clear expectations for quality, consistency, and decision-making.

In production environments, this concept is applied with process controls such as human review, terminology alignment, and repeatable quality checks across multilingual content.

Key Concepts

  • core principle of hallucinations
  • workflow-level implementation
  • terminology and quality consistency
  • human validation before publication

Where It Is Used

  • localisation workflows
  • AI translation pipelines
  • multilingual content production
  • cross-referencing related concepts such as Human Evaluation

Explore Trad AI

Open the workspace